OLego Documentation v1.0.6

What is OLego?

OLego is a program specifically designed for de novo spliced mapping of mRNA-seq reads. OLego adopts a seed-and-extend scheme, and does not rely on a separate external mapper. It achieves high sensitivity of junction detection by strategic searches with very small seeds (12-14 nt), efficiently mapped using Burrows-Wheeler transform (BWT) and FM-index. This also makes it particularly sensitive for discovering small exons. It is implemented in C++ with full support of multiple threading, to allow for fast processing of large-scale data.

OLego is an open source code project and released under GPLv3. The implementation of OLego relies heavily on BWA (version 0.5.9rc1, http://bio-bwa.sourceforge.net/). It also uses some source code from the Jim Kent source code tree (http://genome.ucsc.edu/admin/git.html).

Versions

• v1.0.6 ( 08-09-2012 )
  • Added option –max-multi (default:1000) to avoid huge data in a single line.
  • Added option –num-reads-batch.
  • Fixed a bug in the junction connecting step.
• v1.0.5 ( 07-16-2012 )
  • Minor bug fixed (the old code crashes in a very rare case).
  • v1.0.4 ( 06-12-2012 )
  • Option changes ( do single-anchor search by default now ).
• v1.0.3 ( 06-10-2012 )
  • Now supports strand specific library
  • Fixed bugs about XS
• v1.0.0 ( 05-15-2012 )
  • The initial Public release

Prerequisites

The major programs of OLego ( olego and olegoindex ) can be installed and ran on Unix-based system (Linux or MacOS) with GCC compiler installed. We provided scripts for post analysis and regression model construction, these codes may require Perl and R.

Download

The code and binary files can be found at http://sourceforge.net/projects/ngs-olego/files/ , we are regularly updating the code, so please check regularly to keep your code updated. The code can also be reached via git:

git clone git://git.code.sf.net/p/ngs-olego/code ngs-olego-code

The main programs of OLego (olego and olegoindex ) can be installed and run on Unix-based system with GCC compiler installed. We also provide scripts for post analysis and regression model construction. These codes may require Perl and R installed.

Installation

To compile OLego on your computer, please go to the OLego directory and type:

make

If everything goes right, you will find two executable files olegoindex and olego in the folder.

We also provide binary executable files at http://sourceforge.net/projects/ngs-olego/files/ for x86_64 and i686 Linux systems.

Please feel free to report any problems you come up with.

Usage

Build the index for the genome sequence

To run OLego, you need a BWT index for the reference sequences. We use exactly the same genome index used by BWA in the current version, although this will likely change in the future. For your convience, you can build the index with olegoindex that comes with this package:

olegoindex [-a bwtsw|div|is] [-p STR] <in.fasta>

Arguments:

Argument	Description
<in.fasta>	This is the fasta format file with the reference sequence. Please put all the sequences (from different chromosomes ) in a single file.

Options:

Option	Description
-a	BWT construction algorithm: bwtsw or is [default: bwtsw]
-p	prefix of the index [default: the same as the fasta file name]

Caution: please use “-a bwtsw” for long genome (like human or mouse genome).

There will be 8 files (prefix.pac, prefix.ann, prefix.amb, prefix.rpac, prefix.bwt, prefix.rbwt, prefix.sa, prefix.rsa) generated after olegoindex finishes.

Running OLego

Now you can map your mRNA-seq reads to the genome with olego:

olego [options] <prefix> <in.fastx>

The arguments and options are decribed as below:

Arguments:

Argument	Description
<prefix>	The prefix of the genome sequence index, including the path and the base name.
<in.fastx>	Either fasta or fastq file would work as input.

Basic options:

Argument	Description
-o,–output-file	Name of the output file [ default: stdout ]. This file will be in SAM format, with some customized tags. Please see the details of the file format below.
-j,–junction-file	Annotation file for known exon junctions. It is in BED format and please see the junc format description below.
-n,–non-denovo	No de novo junction search. Note that if junction annotation file is provided by -j, these “known” junctions will still be searched.
-t,–num-threads	Number of threads (INT) [ default: 1 ]. OLego fully supports multiple threading, if you have multiple CPU cores on your computer, please specify the number of CPUs you want to use with this option.
-r,–regression-model	The file with the parameters for the logistic regression model. The mouse model will be used if no file is selected. The model file contains the parameters for the regression model (the coefficients, the PWM and the background ). We have provided model files for mouse and human. User-defined model can also be generated with the regression_model_gen scripts for any species. Please see its usage below.
-M,–max-total-diff	Maximum total difference between query read and reference sequence. Either INT or FLOAT number can be used for this option. An INT number will specify the maximum total edit distance allowed for each alignment. A FLOAT number will specify the fraction of missing alignments given 2% uniform base error rate. This parameter is the same as -n in BWA. [default: a FLOAT number 0.04 ]
-w,–word-size	The size of the seed used in junction search (INT) [ default: 14 ]. The default seed size is recommended for reads >100 nt. For shorter reads, a smaller number can be used. e.g., 12 nt for 36 nt reads. The seeds will be evenly distributed on the read from the start to the end, so please try to cover the read as much as possible with a reasonable seed size. (13 nt for 36 nt reads is a BAD example. )
-W,–max-word-occ	Maximum number of matches of a seed (INT) [ default: 1000]. If a seed has more than this number of hits on the genome, then it will be considerred repeptive and all of its hits will be discarded.
-m,–max-word-diff	Maximum edit distance allowed for each seed (INT) [ default: 0 ]. Since our seed size is smaller than other programs, we recommend that the user use a small number for this option.
-I,–max-intron	Maximum intron size for de novo junction search (INT) [ default: 500000 ].
-i,–min-intron	Minimum intron size for de novo junction search (INT) [ default: 20 ].
-e,–min-exon	Minimum micro-exon size to be searched (INT) [ default: 9 ].
-a,–min-anchor	Minimum anchor size in de novo single-anchor junction searches (INT) [ default: 8 ]. We define “anchor size” as the smaller number of matched nucleotides on the read at the end of the junction.
-k,–known-min-anchor	Minimum anchor size in single-anchor junction searches when the junction is in the annotation file specified by -j (INT) [ default: 5 ].
-v,–verbose	Verbose mode [ default: false ].